Classification of Myoelectric Signal using Spectrogram Based Window Selection
This paper presents a study of the classification of myoelectric signal using spectrogram with different window sizes. The electromyography (EMG) signals of 40 hand movement types are collected from 10 subjects through NinaPro database. By employing spectrogram, the EMG signals are represented in time-frequency representation. Ten features are extracted from spectrogram for performance evaluation. In this study, two classifiers namely support vector machine (SVM) and linear discriminate analysis (LDA) are used to evaluate the performance of spectrogram features in the classification of EMG signals. To determine the best window size in spectrogram, three different Hanning window sizes are examined. The experimental results indicate that by applying spectrogram with optimize window size and LDA, the highest mean classification accuracy of 91.29% is obtained.
How to Cite
Open access licenses
Open Access is by licensing the content with a Creative Commons (CC) license.
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.